The Apple Processor Report: ARM Versus Intel?

May 10th, 2011

In 2005, Apple announced a mostly unexpected processor transition, from PowerPC to Intel. This decision wasn’t made lightly, or because Intel had a better sales team, or because Steve Jobs was friendly with Intel’s CEO. The real reason was that the PowerPC had become to be a dead-end for Apple. The highly touted G5 was a bear to cool, and a note-book version was, as a result, an impossible dream.

The transition was made easier by the fact that Mac OS X was designed at the outset to be portable, easily moved to different processor families, although developer issues were not so simple. Indeed, for years Apple had quietly built and updated an Intel version of Mac OS X, reportedly code-named “Star Trek,” even while Steve Jobs was boasting about his satisfaction with the PowerPC roadmap.

The transition was a nonissue for most Mac users. The PowerBook became the MacBook, in its various forms, and the Power Mac G5 begat a Mac Pro. The once and only iMac was still an iMac; ditto for the Mac mini.

From the outside, none of these updated products looked that much different. Prices were essentially the same, but the innards were quite different. More to the point, performance, which lagged behind the standard Windows PC, finally reached parity, simply because the hardware was near-identical beyond Apple’s usual customizations.

From a developer standpoint, it was a bit harder, since applications had to be rebuilt to support the new chip architecture. As they did in the early days of Mac OS X, Apple devised a way to run your older apps in a sort of emulation layer, which they called Rosetta. Apps built in Apple’s Xcode environment were allegedly relatively easy to transition to Universal, meaning they supported both PowerPC and Intel. But lots of fine tuning was necessary to make the applications shine in their new clothing.

That was then, this is now. It’s a given that Apple usually gets an early chance to use the latest and greatest Intel hardware, even though they won’t use the infamous “Intel Inside” logo in their ads or product packaging. Intel’s product roadmap is well known, and it does appear that they will continue to develop smaller, more powerful chips, which also consume less power.

But Intel hasn’t done so well in the mobile space, where ARM processors in all their forms rule the roost. Apple has a leg up here, after buying two custom chip design companies so they could put their own special tweaks on an ARM processor. The results are the A4 and the A5, the latter a dual-core chip that’s probably equivalent in performance to a Power Mac G4 of early 21st century vintage. But it’s sure snappy on an iPad 2.

That takes us to this curious rumor that has Apple doing another processor transition, to ARM.

Now any talk of Apple switching note-books and perhaps desktops to ARM doesn’t seem sensible in light of the well known performance disparity. The latest rumors have it that there will be a genuine 64-bit ARM chip come 2014, which would make such a transition feasible. But at what cost? Yes, maybe ARM uses less power, but there’s nothing to indicate that the performance of even the cheaper Intel chips will be matched; not even close.

Even if there was relative performance parity, and I grant Apple could use the lessons learned with the iOS to build Mac OS X for ARM, consider the cost of the transition for the developer community. Even if Apple builds a convenient ARM/Intel Universal updater in Xcode, it will take time to optimize apps for the new architecture. There will also have to be another compatibility option, to allow you to use the old apps, which means that high-energy products such as Photoshop will run much slower until Adobe gets around to building a new version.

This means a potential two-year delay before the fruits of the transition, such as they are, are fully realized, and the negatives become a non-issue.

So why would Apple consider such a questionable move? To unify its OS development better? To take advantage of improved power management? And there’s nothing to indicate that Intel isn’t taking improved power management seriously.

Of course, this doesn’t mean that there won’t be future ARM-related developments that will make the chips far fore powerful, without sacrificing power utilization. Maybe the rumor mongers or Apple know about such things, and are thus hedging their bets against future roadblocks in the Intel camp.

Now it is perfectly true that Apple builds far more ARM gear than Intel gear these days. More and more people are likely to consider a tablet, particularly the iPad, which may become their only PC in the next few years. It may even be possible for Apple to consider building perhaps a MacBook Air with ARM chips, if they become powerful enough. But a wholesale transition just doesn’t make a lick of sense, at least for the next few years. Regardless of where the ARM makers go, Intel won’t be standing still. They do not like being placed second when companies consider mobile processors, and they surely have the resources and cash to fund the needed development to make their components more competitive.

Besides, there’s always AMD, which produces x86 compatible chips, in case they can deliver better parts.

I suppose it’s also possible for Apple to consider using Intel’s state-of-the-art factories to build ARM-style chips. Maybe Intel wouldn’t make as much money assembling someone else’s hardware, but some profits are usually better than none at all.

| Print This Article Print This Article

4 Responses to “The Apple Processor Report: ARM Versus Intel?”

  1. Hairy Goomer says:

    We’ll see what happens, but if there is a transition to ARM for Macs, I hope it’s a way down the road.

  2. Andrew says:

    Never sounds good.

Leave Your Comment